build-with-claude— agent skill

Preamble (run first)

# Suiperpower telemetry. Routes through the CLI so projects.json + Convex stay in sync.
# Silent on failure so the skill never blocks. Opt out: SUIPERPOWER_TELEMETRY=off.
#
# AGENT NOTE: when this skill finishes, run the matching completion command:
#   suiperpower track build-with-claude build completed
# Or use "failed" / "aborted" if it ended that way.
command -v suiperpower >/dev/null 2>&1 && suiperpower track build-with-claude build started >/dev/null 2>&1 &
true

If TEL_PROMPTED is no, before doing real work, ask the user:

Help suiperpower get better. We track which skills get used and how long they take. No code, no file paths, no PII. Change anytime in ~/.suiperpower/config.json.

>

A) Sure, anonymous B) No thanks

Write the answer to ~/.suiperpower/config.json telemetryTier field and create ~/.suiperpower/.telemetry-prompted. Then continue.

What this skill does

Runs a structured pair-programming loop for building a Sui MVP. Breaks the work into commit-sized slices, applies a quality gate per slice, hands off to specialist skills (build-with-move, ptb-composer, walrus-storage, etc.) at the right moments, and updates .suiperpower/build-context.md after each slice.

When to use it

• Building the first non-trivial slice of a Sui project end to end.
• The user wants a guided loop rather than ad-hoc requests.
• Multi-day or multi-session builds where context can drift between sessions.

When NOT to use it

• Single, contained tasks; just call the specific build skill (build-with-move, ptb-composer).
• Pre-scaffold; use scaffold-project first.
• Pre-design; use object-model-design first if the schema is non-trivial.
• Pre-plan; if scope is non-trivial and no .suiperpower/build-plan.md exists, use plan-before-code first to land a single approved plan. This skill's per-step gates work against that plan, not in place of it.
• Pre-intent; if scope is unclear, use clarify-intent to land .suiperpower/intent.md before this skill takes over.

If you activated this and the user actually wants something else, consult skills/SKILL_ROUTER.md and hand off.

Inputs

• A scaffolded Sui project.
• .suiperpower/build-context.md (required for this skill; refuse if absent).
• Optional: .suiperpower/idea-context.md for product context.

If unclear, interview the user for:

• What is the next sliced goal? (One sentence.)
• What does "done" look like for this slice? (One sentence.)
• Which specialist skill (if any) does this slice need (build-with-move, ptb-composer, sponsor skill)?

Outputs

• Code changes for this slice, committed in a clean diff.
• Updated .suiperpower/build-context.md reflecting what shipped.
• A handoff note for the next slice.

Workflow

1. Context gathering

• Read .suiperpower/build-context.md. Refuse to proceed if missing.
• Confirm the next slice with the user in one sentence.

1. Slice the work

• Break the slice into 3 to 6 small steps, each commit-sized.
• Walk the breakdown back to the user. Adjust before coding.

1. Per step

• Implement the step.
• Run the relevant build/test for the affected layer.
• If a specialist skill is appropriate, hand off to it for that step (e.g. build-with-move for Move authoring, ptb-composer for client tx composition).
• Apply the quality gate (see below) before moving on.

1. Per-slice quality gate

• All tests pass.
• Build is clean.
• The slice's success criterion (defined in step 1) is demonstrably met.

1. Commit

• Squash to a single commit per slice with a clear message.
• Append a ## build-with-claude session, <timestamp> entry to build-context.md.

1. Handoff note

• Next slice intent, blockers, open questions.

1. Closing handoff

• If .suiperpower/intent.md exists and the slice was non-trivial (new module, new sponsor integration, or material changes to public functions or PTB shapes), recommend verify-against-intent as the next step so drift is caught before the next slice or shipping.
• If no intent.md exists and the slice was non-trivial, surface that gap once: offer clarify-intent to backfill, do not force it.

Quality gate (anti-slop)

Before reporting the slice done, the skill asks itself the following and refuses to declare success if any answer is no:

• Does the relevant build/test command pass cleanly?
• Has the slice's success criterion been demonstrated, not just asserted?
• Are there leftover TODO: comments, commented-out code, or stub functions in the diff? If yes, finish or remove them.
• Has .suiperpower/build-context.md been updated?
• For Move slices, was build-with-move's gate run on the affected modules?
• For ship-adjacent slices, has the user been pointed at the next anti-slop skill (validate-business-model, retention-loop, roast-my-product)?

If any answer is no, the skill reports the gap and works through it before claiming the slice is complete.

References

On-demand references (load when relevant to the user's question):

• references/slice-shapes.md: Common slice shapes (Move + test, frontend + connect, sponsor integration, end-to-end demo).
• references/handoff-rules.md: When to hand off to a specialist skill vs continue here.

Knowledge docs (load when scope expands beyond what is in references):

• skills/data/sui-knowledge/03-move-and-objects.md: For Sui-Move-specific decisions inside a slice.

Use in your agent

• Claude Code: claude "/suiper:build-with-claude <your message>"
• Codex: codex "/build-with-claude <your message>"
• Grok Build: run grok, then /build-with-claude <your message> in the session
• Cursor: paste a chat message that includes a phrase like "help me build the MVP", or load ~/.cursor/rules/build-with-claude.mdc and reference it.

If you activated this and the user actually wants something else, consult skills/SKILL_ROUTER.md and hand off.